Is there ant specific config i need to check, my pbsnodes -a never catch the memory usage on nodes.
Below is a test screeshot, in which I am running jobs on all 4 cores. And the pbsnodes -a shows resources_assigned.mem = 0kb
While free -m and also htop does show some memory is getting used.
Did you request memory in your qsub ?
Please share your job submission scripts along with resource requested via qsub.
Please submit the below job for 5 times and check pbsnodes -a output
qsub -l select=1:ncpus=1:mem=50gb – /bin/sleep 10000
This is not related to the memory used by the job on the compute node. This is what is being assigned to match the (qsub) job requested mem for scheduling.
if a job is submitted with qsub -l select=1:ncpus=1:mem=5gb , then
Oh thanks Adarsh for clarifying. I was not aware of this. I thought it should just display the memory getting used irrespective of what we ask for in qsub.
It works fine now. I tested by adding mem parameter.